RE: Multiple rdf:type elements

On Thu, 6 Jun 2002, Ronald Daniel wrote:

>
> It has been discussed in the past, and is considered
> a feature rather than a bug.

Yes, its a natural consequence of doing typing in a distributed, Web
environment over a single flat space of URI-named objects. If RDF only
allowed a resource to sit in one class hierarchy, it'd force us to do a
*lot* more pre-coordination of RDF vocabulary design.

Something, eg the resource named http://xyz.example.com/abc could
simultaneously be considered a resource of type rss:Item,
edu:OnlineLearningObject, image:Photo and dct:MediaObject. A single
RDF/XML description of it might include properties that apply to various
of those types of thing. If we instead said that a resource is only a
member of one class (and its superclasses, presumably) we'd force
competition amongst complementary resource description activities: we
could only use one of several vocabularies at a time.

In this regard, RDF is pluralist by design rather than accident...

Dan


> Ron
>
> > -----Original Message-----
> > From: Alan Lillich [mailto:alillich@adobe.com]
> > Sent: Thursday, June 06, 2002 11:18 AM
> > To: www-rdf-interest@w3.org
> > Subject: Multiple rdf:type elements
> >
> >
> >
> > An example of an rdf:Description element containing multiple rdf:type
> > elements came up in a recent discussion.  Clearly the syntax
> > allows this,
> > and any one of the rdf:type elements could be used to produce
> > a typedNode
> > form.
> >
> > Has there ever been any discussion about the utility or
> > meaning of this?
> > Perhaps to model multiple inheritance? Is it an oversight in
> > the syntax?
> > Should RDF implementations complain about it if they have a
> > checking mode?
> >
> > Alan Lillich
> > Adobe Systems
> >
>

Received on Thursday, 6 June 2002 17:10:45 UTC